Trends of cervical cancer in Greenland: A 60-year overview

Abstract

Background. Due to its extraordinarily fast economic and social transition, virtually closed borders before 1940 and, moreover, that 85% of the population has the distinctive genetics of the Inuit, Greenland is a very interesting country to study cervical cancer from a historical perspective. Nevertheless, little has been reported about long-term cancer trends in Greenland. Our aim was to describe and interpret the incidence of cervical cancer from 1950 to 2009.

Material and methods. We systematically searched PubMed for articles reporting the incidence of cervical cancer in Greenland. We supplemented this with data for 1980–2009 obtained from the Chief Medical Officer of Greenland.

Results. Incidence of cervical cancer was around 10 per 100 000 women (age-standardised, world population, ASW) in the 1950s, 30 per 100 000 in the 1960s, and in the 1980s around 60 per 100 000. From 1985 onwards, the incidence of cervical cancer started decreasing to the current level of 25 per 100 000.

Conclusion. The steep increase in the incidence of cervical cancer from the 1950s onwards is unlikely to be explained by increasing completeness of data. In parallel with the economic development, however, out-of-wedlock births (proxy for sexual behaviour) increased dramatically from 1935 onwards while tobacco use increased from the 1950s onwards. From the late 1960s to around 1990, data suggested rather stable but high levels of sexual habits. The decrease in the incidence of cervical cancer since 1985 is consistent with the introduction of screening. The data strongly suggested that the increased burden of cervical cancer in Greenlandic women was real and followed earlier changes in sexual behaviour; these changes were likely a consequence of the tremendous societal changes.

Since the Second World War, an extraordinarily fast economic and social transition has taken place in Greenland. Nowadays, Greenland is an autonomous part of the Kingdom of Denmark with a high gross domestic product per capita [1], a healthcare system that is well developed, and cancer registration that has been in place since the 1950s [2–4].

Before 1940, Greenland was a closed country where no one could enter without permission from the Royal Greenland Trading Department [3]. Until the early 1950s, most of the Inuit population lived in small isolated settlements. Thereafter, communication within Greenland, and with the world outside, increased greatly [5,6]. The country also became more urbanised; in 1940, only 15% of the population grew up in towns, and by 1965, this proportion increased to 50% [7]. Also the population increased from 12 000 in 1901 to around 25 000 in the 1950s and to the current population of around 56 000 from 1990 onwards [3,8].

Cervical cancer, a consequence of an infection with human papillomavirus (HPV), is known to be related to social background [9]. On top of this, the Inuit seem to have distinctive genetic traits, and genetic factors are increasingly in the focus of understanding the epidemiology of infectious diseases [10–12]. In the 1970s and the 1980s, Greenland had one of the highest recorded incidence rates of cervical cancer in the world [13,14].

All these factors make Greenland a very interesting country based on which epidemiology of cervical cancer could be studied from a historical perspective. Nevertheless, Greenlandic data have not been included in international historical comparisons, e.g. Cancer incidence in five continents [15]. Hence, the purpose of this article was to describe and interpret the long-term trends in cervical cancer during the period 1950–2009.

Material and methods

We undertook a systematic PubMed search to identify all peer reviewed sources, published until September 2013, reporting cervical cancer incidence rates for Greenland; search strategy “Greenland [MeSH] or Greenland [Title/Abstract]”. Abstracts were read by two authors (BBS, MR) independently. We included only papers that reported incidence rates standardised to the world population (ASW), or data that could be converted to such. The reason we used secondary sources was that with a few cases per year, we found it important to benefit from the checks other authors had made on primary data. All selected publications reported incidence rates for various periods each covering several years. These periods were only partially overlapping when compared across the publications, leading to some variation across publication in the incidence rates.

As an exception, for the period 1980–2009 that was less well covered by the literature, we obtained data on all cervical cancer cases through the office of the Greenlandic Chief Medical Officer. These data were retrieved from the Danish Cancer Registry (Gorm Nørgaard Pedersen, personal communication, 2011). To avoid fluctuations due to small numbers, we analysed these data for 10-year periods. The numerators for each 10-year period were the total numbers of cervical cancers by five-year age groups; the denominators were the total numbers of woman years in five-year age groups, estimated by summing the yearly number of women in the population [8]. The incidence rates were age-standardised using ASW. In order to obtain a smoothed representation of the incidence rate, we computed and plotted the five-year moving average based, for each year, on all available incidence rate estimates (population- and ethnicity-based).

Results

In total, 2748 abstracts were identified based on the search criteria (Appendix, to be found online at http://informahealthcare.com/doi/abs/10.3109/0284186X.2014.883462). Thirty articles were retrieved in full text, among which 22 did not report ASW incidence rates, and were excluded. The data retrieved from the remaining eight articles, all reporting on incidence across the entire age span, were described in Table I.

Table I. Reviewed publications on the incidence of cervical cancer in Greenland identified by a systematic search in PubMed.

Authors and year	Time period	Sources of data on cervical cancer	Definition of a Greenlander (cases)	Definition of population (numerator)	Description of data quality by authors	Diagnosis
a) Nielsen 1986 [4]	1950–1974	A review of all records in Greenlandic hospitals, records on patients sent to Denmark for treatment, records from Danish hospitals, tissue sent to Denmark, deceased persons lists, death certificates and the Danish Cancer Registry	Two parents born in Greenland, or Greenlandic maternal of paternal decent verified from church books, or identified as Greenlandic through pictures, names and information from relatives or hospital personnel	Women born in Greenland and living in Greenland	“Since 1954... many aspects of the Greenlandic health service came close to Danish standards around 1954. Registration of disease in terms of completeness and accuracy was greatly improved”.^a	Only 2 of 101 not histologically confirmed Diagnosis: ICD-O:180/ ICD-7:171
b) Prener et al. 1991 [23]	1953–1985^b	From 1975: Danish Cancer Registry supplemented by pathology reports and death certificates	Born in Greenland and living in Greenland at the time of diagnosis	Women born in Greenland and living in Greenland	“The Greenlandic health system was in a phase of development from 1950–1955”.^c “With introduction of more specialized medical services from 1954, the quality of the health service came close to that of for example Denmark, and registration of diseases and mortality was also greatly improved”.^d	ICD-O codes converted to modified version of ICD-7 (171)
c) Melbye et al. 1984 [17]	1960–1981^b	From 1975: notifications to the Danish Cancer Registry, and pathology reports [92]	From 1975: not further defined [92]	From 1975: not further defined [92]	–	From 1975: not further defined [92]
d) Kjaer et al. 1996 [13]	1969–1988^b	From 1975: Danish Cancer Registry, supplemented by pathology reports and death certificates	Born in Greenland and living in Greenland at the time of diagnosis	Women born in Greenland and living in Greenland	–	Modified version of ICD- 7 converted to ICD-9 (180)
e) Friborg et al. 2003 [18]	1973–1997^b	From 1975: Danish Cancer Registry	Cohort of all women born in Greenland and alive on 1 January 1973 or later^e	Person years at risk in the cohort	–	Modified version of ICD-7 (171)
f) Nielsen et al. 1988 [80]	1976–1985	Pathology reports; cervical cancers were cross-checked with the Danish Cancer Registry	Born in Greenland and living in Greenland	Women born in Greenland and living in Greenland	–	–
g) Nielsen et al. 1993 [82]	1976–1991	Pathology reports; cervical cancers were cross-checked with the Danish Cancer Registry^f	Born in Greenland and living in Greenland	Women born in Greenland and living in Greenland^g	–	–
i) Kelly et al. 2008 [19]	1989–2003	Danish Cancer Registry	Born in Greenland and living in Greenland	Women born in Greenland and living in Greenland	–	ICD-10:53/ ICD-9:180

–, not reported. ^ap. 12–13; ^buntil 1974, the source of information was Nielsen (1986) [4]; ^cp. 11–12; ^dp. 16; ^e The cohort consisted of 72 331 individuals. Follow-up started on 1 January 1973, or at date of birth, whichever occurred last, and ended at the date of cancer diagnosis, death, emigration, registrational disappearance, or 31 December 1997; whichever occurred first. Cancers in Greenlanders living in Denmark at the time of diagnosis were included; ^fThis cross-check perhaps not done for 1989–1991 (Nils Højgaard Nielsen, personal communication, 2012); ^gDenominators for calculation retrieved by authors from Prener et al. (1991) [23] (year 1976) and from Greenland Statistics [8] (years 1977–1991).

Nielsen identified incident cancer cases in Greenland in 1950–1974 using various data sources, including a review of all Greenlandic hospitals records, and all preserved histological material for patients (Table I) [4]. Subsequent publications for this period used his data [13,16–18]. Published data for the period from 1975 onwards combined information from the Danish Cancer Registry, pathology reports and death certificates. Two publications covering 1973–1997 and 1989–2003, respectively, used only data from the Danish Cancer Registry [18,19].

The first certain case of cervical cancer in Greenland was described in 1915 [20], and in total 11 cases were reported until 1949 [4]. The incidence of cervical cancer was around 10 per 100 000 in the 1950s (Figure 1). It then increased to around 30 per 100 000 in the 1960s, and to around 60 per 100 000 in 1970–1974. It remained at that level until the mid-1980s. Thereafter, the incidence started to decrease, first to around 50 per 100 000 in 1985–1990, and then to around 25 per 100 000 in 2000–2009. Although variation in the incidence rates was seen by publication, they pointed to the same trend.

Figure 1. Incidence of cervical cancer in Denmark and Greenland. ^a For Denmark, the source was NORDCAN and a three-year smoothing was used [21]. See Table I for Greenland: a) Nielsen 1986 [4], b) Prener et al. 1991 [23], c) Melbye et al. 1984 [17], d) Kjaer et al. 1996 [13], e) Friborg et al. 2003 [18], f) Nielsen et al. 1988 [80], g) Nielsen et al. 1993 [82], h) Gorm Nørgaard Pedersen, personal communication, 2011, i) Kelly et al. 2008 [19]; ^b For a smooth representation of the incidence rate, we computed and plotted the five-year moving average based, for each year, on all available incidence rate estimates (population- and ethnicity-based).

There seems to have been a trend toward a slightly increasing age at diagnosis from the 1950s/1960s towards the following decades [4,13] but due to the small number of cases the estimates are imprecise. For the more recent period (Gorm Nørgaard Pedersen, personal communication, 2011), the average age of diagnosis was 43 years in the 1980s, 43 years in the 1990s and 41 years in the 2000s. In the 1980s most cases were diagnosed in the age-group 35–44 whereas this was 40–49 years in the 1990s and 30–39 years in the 2000s.

Discussion

General findings

The pattern of cervical cancer in Greenland has differed from that of several other Nordic countries [21]. In Greenland, the incidence increased sharply from a low level in the 1950s to a very high level in 1970–1985. In Iceland (another historically isolated nation) a marked increase from initially a low level, was also seen from the 1950s to the 1960s [21]. In other Nordic countries, such as Denmark, the incidence was relatively stable but higher than in Greenland until the mid-1960s (Figure 1). Although Denmark is often given as an example of a country with a high incidence rate of cervical cancer, it should be emphasised that the peak incidence in Greenland around mid-1980s was about 2–3 times higher than the peak incidence in Denmark around mid-1960s. Since the peak levels, however, the incidence rates have become substantially lower in both Greenland and Denmark, but the incidence in Greenland has remained higher than in Denmark.

Interestingly, also for the Canadian Inuit a sharp increase in the number of cervical cancers has been suggested for the period 1950–1980 [22]. In Alaskan natives (50% Inuit), a three-fold increase in the incidence of cervical cancer was observed from 1969–1973 to the peak in 1979–1983 [13].

The low incidence in Greenland during the 1950s and the 1960s, and the later increase to one of the highest ever recorded levels in the world raises a question whether this pattern is an artefact, owing to an improved completeness of data sources over time, or it is a consequence of a real change in risk factors. This is investigated below.

Evidence indicating incomplete registration of cervical cancer cases

An almost 100% increase in the incidence of all cancers in Greenland from the 1950s to the 1960s [4] has been seen as a sign of general underreporting in the 1950s [23–25]. Little has, however, been published on the quality of cancer registration in Greenland [24].

In Greenland, settlements are scattered and relatively small. Particularly in the early period covered by our study, this posed a challenge for healthcare delivery [26]. Furthermore, in 1950, hospitals were poorly equipped, and only 15 physicians practiced in the whole country [3]. This may be pointing to incomplete recognition of the relevant cancer cases by the healthcare system.

Evidence against incomplete registration of cervical cancer cases in the early study period

Already in 1935, it was described to be easy to motivate patients for hospitalisation [27], and doctors working in Greenland described the patients’ willingness to seek medical help [28,29]. A modern healthcare service administered by the Danish state was launched in 1950 [23], with health care and medicines being free of charge ever since [2,23,30]. Specialised Danish doctors started travelling the districts in the summer, and many Greenlandic patients were flown to Denmark for treatment [4]. Also, Greenlandic histological specimens started being examined in increasing numbers in Denmark. Hence, recognition of the relevant cancer cases by the healthcare system was improving already in the 1950s.

From 1 January 1967, death certification, a source of information for cancer registration, became mandatory, although also in preceding years most hospital deaths were certified [4]. A study of 282 deaths in Greenland in 1968–1981, found a sensitivity of 1 for the diagnosis of cancer as cause of death, and the positive predictive value was 0.78 [31]. The above speaks against underreporting, especially in the period following the 1950s. Also, all sources covering the 1950s and 1960s supplemented data from the Danish Cancer registry with other sources such as the death certificates and review of hospital records (Table I). The suspected underreporting, due to the 100% increase in the incidence of all cancers in Greenland from the 1950s to the 1960s, is probably less relevant when it comes to cervical cancer, since it is a cancer that does not demand advanced equipment to diagnose. In addition, the proportion of cervical cancers of all female cancer cases increased from 11% in the 1950s to 20% in the 1960s [4], suggesting a real increase in the incidence of cervical cancer.

From 1963–1985, a very small proportion of uterine cancer cases was classified as “not otherwise specified”, but in 1953–1962, the proportion was higher [23]. Including these cases as cervical cancer cases would increase the early incidence somewhat, but it would not change the overall pattern.

Evidence on changes in sexual behaviour

Human papillomavirus (HPV) prevalence

HPV DNA prevalence in Greenland was first assessed in the 1980s [32,33]. Surprisingly, the prevalence among 20–39-year-old women was lower in Greenland than in Denmark [34]. This unexpected finding has since been explained as a consequence of an earlier sexual debut in Greenland, and, accordingly, a higher HPV sero-positivity in Greenland than in Denmark [13,35]. Lacking HPV prevalence data for most of the 1950–2009 period, we will instead discuss data on sexual behaviour, a proxy risk factor for cervical cancer.

Sexual behaviour

For the first half of the 20th century, only case reports on sexual behaviour [36] and sexually transmitted diseases (STDs) exist [37,38]. In 1937, Bertelsen (and two other sources later [39,40]) described that the Greenlandic population had a straightforward relation to sex [27]. Multiple sexual partnerships were not unheard of [36,41] although the often referred to story of Inuit men sharing their wives with strangers in the early isolated days appeared not to have been a (common) reality in Greenland [39,42]. The proportion of out-of- wedlock births was around 5% of all births from 1850 (when Greenland became entirely Christianised [3]) to 1935 [43], but then increased to 20% in 1950, and to 37% in 1965 [43]. From 1937 to 1968, the average age at marriage increased from 22.5 to 25 years [36], but this could unlikely fully explain the dramatic increase in out-of-wedlock births. For comparison, in Denmark the proportion of out-of-wedlock births remained at about 10% throughout 1900–1965 [43]. Before 1930, only 18% of Greenlandic women gave birth within the first nine months of marriage, whereas this was 48% in Denmark, indicating that the difference in out-of-wedlock births was probably not due to unmarried pregnant women in Greenland more often getting married compared to Danish women [36]. Hence, it seems that the increase in the proportion of births outside of marriage in the period 1935–1965 most likely reflected a change in sexual activity. This change, seen around 15 years before the increase in cervical cancer cases, was consistent with the known time frame of cervical cancer aetiology [44,45]. Thereafter, intra uterine devices were introduced [36,46], making out-of-wedlock births no longer a usable proxy for sexual behaviour [43].

Among girls interviewed at age 15–19 years in 1966–1968, 58% had initiated sex by age 15 years with 55% reporting more than five partners, and 34% more than 10 partners [36]. Twenty years later, women from Nuuk aged 20–39 years reported that about 15% had experienced first intercourse by age 13 years, and 85% by age 16 years [32,34]. Among these women, 53% and 61%, respectively, reported 20 or more sexual partners [32,33], with 22% reporting more than 40 [47]. For women of the same age from Copenhagen, the corresponding numbers were 15% with ≥ 20 sexual partners, and 4% with ≥ 40. These two studies and others [35,48–53] indicated that a high number of sexual partners and early sexual debut continued to be common in the 1980s and 1990s. This is also supported by a high level of induced abortions, around 60 per 1000 in 1986 [54] (about four times the Danish level [55]), and by a high incidence of chlamydia, 2500 per 100 000 female population in 1995, (6–7 times of the Danish level) [56]. A survey of the adult Greenlandic population from 1993/1994, though, found the number of sexual partners to be somewhat different: during the year before the survey, 14% reported no sexual partner, about 50% had one partner, 17% had 2–4 partners, and 22% had five or more partners [57]. However, the response rate was only 36%.

Olsen [43], and later also From [58], suggested that the reported changes in sexual behaviour could be seen as an indicator of difficulties in adaptation to the very rapid economic and social development in the 1940s–1960s. (Suicides and violence have been described as additional indicators [3,5,59]). As a result of the economical and societal changes, a feeling of safety was lacking. Resorting to sexual activity could be seen as a substitute for this feeling of safety [28,36]. Olsen also pointed to the importance of the influx of (socially attractive) young Danish men in the 1950s and the 1960s, putting sexual pressure on women [36,41,43]. A study of Danes showed that travellers to Greenland had 2–3 times more sexual partners, prostitute contacts and STDs than Danes not travelling to Greenland [60]. It has furthermore been suggested that the sexual behaviour of women in the 1950s and 1960s could have been influenced by the historic Inuit sexual morale that Berthelsen described in 1937: that sex was considered necessary, pleasant and invigorating [27,40].

Other factors potentially increasing the risk of cervical cancer

Genetic/immunological factors

Cervical cancer has been suggested to have a genetic component [61–63]. The Inuit gene pool is in many respects similar to that of East Asian populations. Furthermore, the isolation of the circumpolar region resulted in a unique genetic situation where the Inuit lived in almost complete isolation for 5000 years [64].

Historically, the Greenlandic Inuit have had an excess risk of three virus-related cancers, nasopharyngeal cancer (where it is widely accepted that a genetic factor is involved [65]), salivary gland cancer, and cervical cancer [17]. This was confirmed in a more recent study comparing Inuit from the US, Canada, and Greenland to US Whites [66]. The risk was 50 times higher for nasopharyngeal cancer, and for salivary gland and cervix cancer it was 10 and three times higher, respectively [66]. For first-degree relatives of Greenlandic Inuit with nasopharyngeal cancer, an eight-fold increased risk of salivary gland cancer and a doubled risk of cervix cancer were found compared to non-relatives [67]. Two studies also observed a persistently increased incidence of the three cancers among the Greenlandic Inuit migrating to Denmark [16,68], whereas for Danes who worked in Greenland, the risk was not increased [69]. In Inuit populations, furthermore, excess risk has been detected for tuberculosis and pneumococcal infections [70–72]. Correspondingly, in 1995 children in Nuuk were prescribed 2–3 times more anti-infectives than children in Sweden, Norway or USA [73]. This was observed even though recommendations regarding drug use in Greenland are based on Danish traditions, which are similar to those in other Scandinavian countries.

The genetic component of cervical cancer is possibly related to the immune system, which is important in the pathogenesis of HPV-induced diseases [74]. Deficient natural resistance to particular pathogens can be seen in populations only exposed to those pathogens for a short time period [75]. For Greenland it could have been the case that HPV was not spread around the country until after im(migration) became more common in the 1950s, resulting in an increased background risk of cervical cancer.

Tobacco use

Smoking increases the risk of cervical cancer [76]. The average daily consumption of cigarettes in Greenland increased strongly from the 1950s to the 1980s [3], when 90% of women aged 20–39 were smokers [34]. This appears to be parallel with the increase of cervical cancer from the 1960s to the 1980s [23]. The consumption of cigarettes decreased after the 1980s [3,23,77].

Protective factors for cervical cancer

Intra uterine devices, for which some protective effect for cervical cancer has recently been suggested [78], were very commonly used by 1970 [43]. Newer information on their use is sparse, but a decreasing trend was suggested [79].

Smear taking started in Greenland in the late 1960s [80–82]. From 1976 to 1985, screening coverage among women aged 20–59 years was below 50% [80,81]. In 1976–1985 and 1986–1991, 65% and 57%, respectively, of cervical cancer patients had never been screened, and 14% and 23%, respectively, had inadequate follow-up of screen-detected abnormalities [80,81]. A poor quality of smears was also suggested for that period [82]. A national organised screening programme for women aged 20–70 years was launched in 1998 [83]. This was initially successful, achieving a coverage of 70%, but experienced problems with the invitational system. The coverage decreased back to 40% in 2007 [24, 83]. Improvements in the invitational system were introduced in 2010 [24,83]. Despite these problems, the decrease in the incidence of cervical cancer from 1985 onwards could be consistent with the introduction of screening. The high use of intra uterine devices around 1970, and the suggested lower levels of tobacco use, may have contributed to this trend.

HPV vaccination was introduced in 2008 [84], and might further reduce the burden of cervical cancer. Vaccination coverage data, however, are not available [85,86].

Strengths of the analysis

Being able to work with cancer registration data for a 60-year period is a possibility only for very few countries. For Greenland, this is the first time that data on cervical cancer incidence have been combined for such a long period. This gave us an opportunity to describe and interpret the epidemiology of cervical cancer in this country in a more comprehensive way than it would have been possible looking only at sub-periods.

For the early period, we included only data from peer-reviewed sources, to ensure a high quality of the data. We furthermore assessed also some non-peer-reviewed reports [2,87–89]. Overall, these reports were in line with peer-reviewed sources.

Weaknesses of the analysis

Most data in the reviewed papers were reported for the ethnicity proxy “Greenland-born”, however, this was not defined in the same way in all publications [90] (Table I). The recent data from the Chief Medical Officer were population based. The difference is not negligible. In 1975, 20% of the population was born outside of Greenland [3,8]. Most non-Greenland born population living in Greenland are Danes [23], and the incidence of cervical cancer has been lower in Denmark than Greenland since 1965 [21]. The population-based incidence in 1980–2009 could therefore be expected to be somewhat lower than the Inuit-based incidence, and part of the decrease from 1985 and onwards could be a methodological artefact. Nevertheless, the Inuit- and population-based estimates were very similar (Figure 1).

Until the late 1980s, a proportion of Greenlandic cancer cases identified by Copenhagen University Hospital, Rigshospitalet (the most comprehensive source since all patients from Greenland have been treated there), were not reported to the Cancer Registry (Nils Højgaard Nielsen, personal communication, 2012). In 1978, 28 (25%) cases were found to be missing in the period 1957–1974. Of the non-reported cases, 13 were from the period 1970–1974 [14]. The cases identified until 1988 were, though, subsequently included in the Cancer Registry (Nils Højgaard Nielsen, personal communication, 2012), but for the period thereafter, some underreporting of Greenlandic cases cannot be entirely ruled out. For head and neck cancers diagnosed in Greenland in 1994–2003, it has been estimated that 96% were included in the Danish Cancer Registry [91]. Hence, incidence estimates from the sources using exclusively the Cancer Registry, including our own analysis of the Registry's data for 1980–2009, could be on the low side also for this reason.

In conclusion, the incidence of cervical cancer in Greenland during the last 60 years is rather extraordinary; particularly the major increase between 1950 and ca. 1985 was unlike the contemporary trends in many other countries. The most probable explanation for the increase is a changed sexual behaviour, in itself likely a consequence of the social changes that the country experienced since the mid-20th century.

Acknowledgements

Bente Braad Sander holds a PhD-scholarship from the Danish Research and Innovation Agency, and has received a grant from Aase and Ejnar Danielsens foundation. Matejka Rebolj is supported by a grant from the Danish Strategic Research Council. The authors would like to thank Nils Højgaard Nielsen for his valuable comments on an earlier draft of this paper, Gorm Nørgaard Pedersen for the data, and Pierre-Antoine Dugué for help with graphic presentation and statistical analysis.

Declaration of interest: Elsebeth Lynge and Matejka Rebolj are currently undertaking a comparative study of new-generation HPV tests, involving collaboration with Roche Diagnostics, Genomica, Qiagen, and Hologic/Gen-Probe. Elsebeth Lynge has served as unpaid scientific advisor to Gen-Probe and Norchip, and Matejka Rebolj's employer has received honoraria for lectures from Qiagen on her behalf. Concerning the present paper, there has been no collaboration with, or support from any of the companies.